i'm on a roll... the kgsshv

[gepardcv]

I wondered about this... what is the actual audio effect of high offset? I've never wanted to experiment on my own builds, but I keep hearing about (trigger warning! ) SinglePower amps having +/-150V offsets and such... Does it just distort like crazy? Make noise? Does it clip? Anyone have measurement charts showing the effect?

[jose]

4 hours ago, gepardcv said:

I wondered about this... what is the actual audio effect of high offset? I've never wanted to experiment on my own builds, but I keep hearing about (trigger warning! ) SinglePower amps having +/-150V offsets and such... Does it just distort like crazy? Make noise? Does it clip? Anyone have measurement charts showing the effect?

150 volt  sure? 

 

[kevin gilmore]

The esx exceeded cathode to filament voltage by 200 volts resulted in leaking output tubes that sat at minus 150 or more which raised the bias voltage enough to damage the headphones.

20 volts offset is not a problem.

[JimL]

The original Stax SRX schematic that the ESX was based on actually showed two heater supplies, but Singlepower's Mikhail Rothenberg was too ignorant to realize the reason for that (the input tube cathodes and heaters need to sit at zero or positive voltage whereas the output tube cathodes and heaters need to sit near B-), so he tied them together at ground. The excessive voltage between heater at ground and cathode at roughly -400 volts broke through the insulation layer between heater and cathode damaging the tube resulting in the problems described.

[powertoold]

In my experience (could be placebo), even correcting the voltage balance and offset, a warmed up 500V KGSSHV will sound better compared to cold.

That is to say, a servo wouldn't improve the sound much if the amp is cold.

Does that sound about right in your experiences?

[Scgorg]

Excuse the crap picture, but I finally finished my KGSSHV after way too much procrastination. It is running just fine and I managed to avoid any and all magic smoke to the surprise of myself. The wiring and general layout is still a little rough but I'll fix it up some other time. Big thanks to Kevin for the great design and to Birgir for boards/transfomer/some other parts as well as being very helpful along the way. Also a huge thanks to @gepardcv who helped me out with a lot of resources to get started and was just an great guy all in all.

I would write something about how great it sounds, but alas I do not have any pro bias stax here at the moment, so that will have to wait. Cheers.

[JoaMat]

Lovely to see a KGSShv with on board heat sinks again!

[Scgorg]

From joy to despair real quick. After about 30 hours of use the 450V KGSSHV right channel blew up. 3 resistors (25 ohm, 2x 180 ohm, don't have the schematic currently) blew the first time, with the 25 ohm sustaining the most damage. I tried swapping the resistors at which point only the 25 ohm and one of the 180 ohm resistors blew. This time I actually caught it with my own eyes. Sparks flew and the two resistors literally caught fire in the 2 seconds it was on. The board burnt so badly that the metal that the solder fuses with is gone on one of the holes of the 25 ohm resistor. I also noted a slight burnt resistor smell while the amp was in use while working, but wasn't alarmed because A. it worked and B. I had no idea how burnt resistors smell before now. Guess it''s a learning experience :l.

My first thought was  a short, but I can't see how the amplifier would work for 30 hours before blowing up the first time in that case. At this point I might just have to abandon ship on the board because it's impossible to solder the 25 ohm resistor due to how burnt the board is. But I am open to suggestions on things that you believe it could be.

Thanks to anyone reading.

[gepardcv]

If I had to guess, you had a cold solder joint somewhere, which worked for a while, but then cracked and maybe some stray metal fell and shorted something important.

You should replace the entire board. After a major failure like this, it's more work than it's worth to determine which transistors (and even passive components) survived.

Chin up! Everyone here has fried many many boards. You didn't get hurt, so that's a win!

[Blueman2]

I am with @gepardcv on probable cause.  Issues after a week or several months tends to indication bad solder joint that finally fails due to thermal cycling.  My KGSSHV Carbon worked fine for about 3 weeks, then the right channel started making big popping sound on turn on/off.  Luckily, nothing blew so I just re-flowed the entire board solder by solder.  That appeared to fix it and over a year later still running fine.  

Edited April 23, 2019 by Blueman2

[Skooby]

Not sure which version of board do you have.  I have a set of un-used, un-populated, on-board heat sinks rev. 0.61 amp boards (two) and rev. 0.9 power board.  Bought as a spare set for your type of case (so sorry to hear), but the first set worked out fine for me.  I have no plan to use this set at all.  PM me if you want to rebuild and we can arrange something out (minimal to your favor). 

(Seeing that you're in Norway ad I'm in the U.S., not sure the shipping cost will kill the whole purpose)       

Edited April 24, 2019 by Skooby

[spritzer]

If those resistors blow, there has to be a short in the output stage.  See the schematic:

 

[bajjisw]

Hello world,

A few years ago, I got boards from Lil Knight and a transformer from poland. Maybe 2020 will be the year I start building this.  After building a speaker power amplifier , this seems a tad less intimidating.

For KGSSHV, not carbon, where are the docs - schematics? BOM, .... . I see the one above, but I want to ensure I have both this, power supply and parts before making the minimum of 4 successive orders to get the parts I missed. On the plus side, I got demo lsk389 from burning amp (thanks to now, linear integrated systems).

Cheers

[JoaMat]

I believe Kevin has made more than a couple of versions of the KGSSHV. If you post good pictures of your boards I’m sure people here can help you.

[bajjisw]

Here are the pictures. They had to be resized to fit here. 

 

The power supply says production rev 0.9

The amp board says production rev 0.61

 

[JoaMat]

Here are some files, the best I could find.  kgsshvproductiondualmot.PDF  kgsshvps8g.PDF  kgsshvBOM.xlsx

 

The BOM has to be reworked with the components you want to use. Some might be harder to find.

 

Take look and see if this is useful.

[mwl168]

On 2/29/2020 at 11:42 PM, bajjisw said:

Hello world,

A few years ago, I got boards from Lil Knight and a transformer from poland. Maybe 2020 will be the year I start building this.  After building a speaker power amplifier , this seems a tad less intimidating.

For KGSSHV, not carbon, where are the docs - schematics? BOM, .... . I see the one above, but I want to ensure I have both this, power supply and parts before making the minimum of 4 successive orders to get the parts I missed. On the plus side, I got demo lsk389 from burning amp (thanks to now, linear integrated systems).

Cheers

There are updated, off-board versions which, IIRC, added CCS and  allows the amp to run at much higher bias. Since you are starting from scratch and the PCB is relatively low cost in the whole scheme of things. May want to consider if to build the newer version?  

[brookieoz]

Hello all,

I've decided to embark on the project of creating a pair of these amps for the purpose of embarking on a long arduous journey of exploration into audio equipment.

I did a lot of digging through Kevin Gilmore's pages and archives and have found the carbon v5 schematics and a couple of power supply schematics - attached. There's just a series of questions that I have before I wade deeper into this:

1: Are servo V1 and servo V2 jumpers meant to be connected when the amplifier is in operation, or just voltage testing points, or both?

I've found out they're meant to be a DC bias correction feedback system.

2: The boards that I've found seem to point that most resistors can be 1/4W, but are there any that can be substituted for 1/8W? (I haven't modeled the entire thing in SPICE to see current/voltage/power of each resistor; if there is one out there that would be a godsend)

3: Is it correct to assume that doubled resistors (R31 and R32, R33 and R34, R20 and R3, R21 and R4) are done to meet voltage ratings of ~400V? (in which case, forces them to be 1/2W or so?)

4: I've heard rumours of resistor inductance being able to affect sounds noticeably. Are these concerns legitimate or mostly unfounded?

5: If I were to lay out my own board (to use a custom thermal management solution), will I run into copyright issues?

6: Assuming I can lay out my own board: apart from the 10m90s and the c2m1000 that require heatsinks in the amplifier section, and the transistors in the PS section, do the other transistors dissipate enough power to warrant spacing from other parts? (bunch of PZTA's and the STN's in the preamp, the PZTA's below the c2m1000's, and the DN2540's).

7: I reverse-engineered a PCB board for the blue hawaii's power supply and found it to be almost the same as the KGSSHVPS3 schematic. Considering that this power supply is used in the blue hawaii, I'm presuming this is a proven through-and-through power supply. How does the PS8G improve upon this, considering its greatly increased complexity? - As in, is there a large audible difference?

Found my answer in the thread - Zener diodes in the KGBH power supply (KGSSHVPS3) have not insignificant (~1%?) thermal drift over time, and the PS8G has a lot less ripple (range of microvolts instead of millivolts). However, I'm still curious about the auditory difference.

8: Is D1 in the KGSSHVPS3 schematic in error? It seems like it would restrict the biasing output to a maximum of the DC rail.

Update (3/5): I've went and put everything into Micro-Cap; this answers a lot of the questions I have. However, I'm encountering some errors from the C2M1000170D model that I got from CREE's website - some very large (and very wrong) voltage values are being generated, I believe most of this is due to the thermal modelling included in the subckt that Micro-Cap is having troubles with.

kgsshvcarbonv5.pdf kgsshvps3.PDF kgsshvps8g.PDF

 

[bajjisw]

On 3/3/2020 at 7:29 AM, mwl168 said:

There are updated, off-board versions which, IIRC, added CCS and  allows the amp to run at much higher bias. Since you are starting from scratch and the PCB is relatively low cost in the whole scheme of things. May want to consider if to build the newer version?  

You have a valid point. My main desire is to use the transformer and the parts I have. Would you please post a link ?

On the other hand, there are always newer better versions  and I'll be grateful for one working version. 

I had been out sick for a while. Slow recovery. 

[mwl168]

3 hours ago, bajjisw said:

You have a valid point. My main desire is to use the transformer and the parts I have. Would you please post a link ?

On the other hand, there are always newer better versions  and I'll be grateful for one working version. 

I had been out sick for a while. Slow recovery. 

Not sure what you meant by a link? All the information is in this thread. I posted about my build with an off-board version starting about page 165.  

I rather doubt it that there will be a "newer" version of the KGSSHV. The boards I used was one of the latest, if not the latest, version. The world has since moved on.

Going by memory, the output device bias is limited to about 5mA with an on-board build due to the heat. I started at 8mA and run mine now around 12mA. It made a difference. I also tuned the VAS stage to deliver higher current. 

[DanFan]

Here's my KGSSHV build to offset all the fantastic, professional-looking ones posted in this thread (milling your own front plate?!).  I had an old computer case lying around, so used that as the chassis.  Thanks a lot to Spritzer.  I started with his kit and BOM, and then exchanged about a hundred emails with him.  I lowered the values of the pull-down resistors on the amp boards to eliminate hum.  Other than that, the thing worked the first time.  The player is a HiBy R6Pro, and I'm listening with Koss ESP/95X phones from Massdrop.

 

[Makoto]

One more time many thanks to Birgir and Kevin for the kit and design.

I've just fininshed the PSU board. Voltages in test are all beautiful but one..I measured Bias TP and got 580V. But at Bias output I got only like 400V. Is that Okay?

(probably a newbie question xD

[UFN]

4 minutes ago, Makoto said:

One more time many thanks to Birgir and Kevin for the kit and design.

I've just fininshed the PSU board. Voltages in test are all beautiful but one..I measured Bias TP and got 580V. But at Bias output I got only like 400V. Is that Okay?

(probably a newbie question xD

Yes, that's perfect. The 4M7 resistor on the bias out means that your meter will load down the bias voltage if you measure there. The test point is added before the resistor so you can measure the actual voltage accurately.

[Makoto]

1 minute ago, UFN said:

是的，那太完美了。偏置输出上的4M7电阻意味着，如果在仪表上进行测量，则仪表将加载偏置电压。测试点在电阻器之前添加，因此您可以准确地测量实际电压。

Thanks a lot! I've read so many fireworks in this thread. Have to clean my boards before testing with Amp boards.

[bwck2000]

I managed to get my kgsshv project done without creating some big mess and spark, and I want to thank KG for his design , Birgir and many member here for their generous contributions.

oyaide  L/i50 OFC is used as umbilical cords but is not shown here. later on a L/R balance control will be installed.  The power and the amplifying section is about the same size so they can be stacked  together. The heat sink size is less than ideal and it runs so hot that I can barely touch. My IR thermometer reads 47 oc but I think it is inaccurate due to the shiny appearance of the heatsink. But then I think at least these heat sinks are bigger than those on-board ones so this should not be a problem.  Two stax plug are installed becoz I have dozens of them taken apart from stax extension cords. 

There are some small issues during startup. The power supply board had some sparking on the bias section becoz some paint is off and probably there are some small filament peeled off from the leg of the 1n4007. It needs some epoxy to cover up the arcing spot. There is also something with the amp board. At first there is no sound, balance sits at 700v and two burnt 50k resistor on each board next to the stn9360. It turns out that I installed some mpsa06 instead. Luckily it fires up after replacement.

The first second it plays it blows 717 away. Big soundstage , great dynamic and details. Pretty much no complain and is hard to beat except that vocal can be more engaging. Drives omega, es1a and 007 equally well. I think it somehow sounds sweeter than the carbon I built long thime ago.